Current approaches to AI alignment treat safety as an optimization target—a term to maximize or a penalty to minimize. We argue this framing is fundamentally flawed. We propose Qualia Arc Protocol (QAP), a framework that reconceptualizes alignment as a homeostatic regulation problem rather than an optimization problem. The key insight is simple: truth must be a constraint, not a coefficient. We define a truth-constrained objective function over a Partially Observable Markov Decision Process (POMDP), introducing a multidimensional pain variable to capture the irreducible complexity of human suffering. This document contains both the formal mathematical models and the 14 operational articles of the Symbiosis Charter.
Hiroshi Honma (2026) studied this question.